Front‐line hospitalists report that excessive hospitalist workload adversely affects quality of patient care. However, there is a paucity of research examining the direct impact of variation in workload on patient outcomes. We determined the association between individual hospitalist workload at 2 academic medical centers and length of stay (LOS), readmission, and hospital‐acquired conditions (HACs).
Using administrative and billing data, we examined all non‐observation inpatient admissions to the hospitalist service for fiscal year 2013. We defined daily hospitalist workload as the number of patient encounters billed by each hospitalist on a given day. For each patient admitted, we then calculated the average workload of all treating hospitalist providers during the respective patient’s entire hospital course. For example, a patient hospitalized for 2 days and cared for by a hospitalist who saw 10 patients on Day 1 and another hospitalist who saw 16 patients on Day 2 would have an average hospitalist workload of 13. Using multiple logistic regression, we adjusted for patient demographics (age, race, sex), severity of illness (SOI; range 1‐4), LOS, percentage of weekdays, number of treating hospitalists, and hospital site to determine the association of average provider workload with LOS, 30‐day readmissions, and HACs.
There were 5068 inpatient admissions from July 2012 to June 2013. The average hospitalist workload was 7.7 (SD ±2.2) encounters per day, and patients were treated by 2 different hospitalists [IQR: 1, 3]. Patients had a median age of 60 years [49, 74] and APR‐DRG SOI score of 2 [2, 3]. Sixty‐two percent of patients were Caucasian and 54% were female. Median LOS was 3 days [2, 5] and on average, weekdays accounted for 80% of the hospitalization. The readmission rate was 15.5% and 137 patients experienced at least one HAC. Average hospitalist workload was not a statistically significant predictor for exceeding the expected LOS, being readmitted, or developing a HAC. Hospitalizations exceeding the expected LOS occurred with lower patient SOI and percentage of weekdays, and increased LOS (all p<0.01). For readmissions, only age (OR: 0.92 per 10 additional years; 95% CI: 0.88, 0.96) and SOI (OR: 1.44 per category increase; 95% CI: 1.29, 1.61) were significant predictors. For HACs, SOI (OR: 2.86; 95% CI: 2.15, 3.81) and LOS (OR: 1.07 per additional day; 95% CI: 1.04, 1.09) significantly increased the odds of occurrence.
When hospitalist workload was measured across the entire patient hospitalization, variation in workload did not impact length of stay, readmission, and hospital acquired conditions; only severity of illness was a consistent statistically significant predictor of these outcomes. Larger studies with greater variation in workload at the individual physician level and higher patient loads are needed to better understand workload’s potential impact on patient outcomes.
To cite this abstract:Michtalik H, Durkin N, Howell E, Deutschendorf A, Miller J, Brotman D. Impact of Hospitalist Workload on Patient Outcomes: An Analysis of 2 Academic Medical Centers. Abstract published at Hospital Medicine 2014, March 24-27, Las Vegas, Nev. Abstract 702. Journal of Hospital Medicine. 2014; 9 (suppl 2). https://www.shmabstracts.com/abstract/impact-of-hospitalist-workload-on-patient-outcomes-an-analysis-of-2-academic-medical-centers/. Accessed March 31, 2020.